Ralph Charles Merkle — SECRECY, AUTHENTICATION, AND PUBLIC KEY SYSTEMS, published in 1979

The precursor to proof of work.

“To solve a puzzle, we must cryptanalyze the cryptogram. Once it is done, we learn the information that was “enpuzzled”, the plaintext of the cryptogram. Just as we can encrypt plaintext to produce a cryptogram, so we can enpuzzle information to produce a puzzle.A puzzle, though, is meant to be solved, while ideally, a cryptogram cannot be cryptanalyzed. To solve a puzzle, all you need do is put in the required amount of effort. In cryptanalyzing an encrypted message, the cryptanalyst relies on redundancy in the message to indicate when the proper key is tried.If the information we enpuzzle is random, there will be no redundancy, and thus no way of solving the puzzle. We must deliberately introduce redundancy into our puzzle, so that it can be solved.”